AX ANALYSIS OF THE LEARXIXG PROCESS IX THE SXAIL 39 



this again may be attributed to the unwonted foot-stimulus 

 and may be accepted as evidence of its effectiveness. 



Notable in the training series of trials of Snails 4 and 5 is the 

 increase and subsequent decline in the average number of reac- 

 tions per response after the initial inhibition. The average 

 number of reactions per ten trials rises to a maximum in the 

 neighborhood of the 150th trial and then declines. An examina- 

 tion of the graphs for these two snails (figs. 2, 3, 8 and 9, Plates 

 II-III) suggests that the training has produced its maximum 

 effect in the neighborhood of the 150th trial. If this were true 

 a longer response to pressure only should be obtained from snails 

 that had been trained for 150 trials than from those of longer 

 training and this response might persist for a longer period. 

 Whether this is actually the case could only be determined by 

 comparing the response to pressure alone in a considerable num- 

 ber of snails with training periods of different lengths. A com- 

 parison of the graphs for Snails 4 and 5 with those for 6, 7, 8 

 and 14 shows that the latter are quite irregular. They show 

 numerous maxima one of which commonly lies in the region of 

 trial 150, but they do not support the view of a maximum train- 

 ing effect at a particular region in the series of trials. Thus an 

 interpretation that seems wholly tenable from a study of two 

 snails becomes untenable when the study is extended to a larger 

 number of individuals. The maxima in the graphs for Snails 4 

 and 5 probably have their explanation in some undiscovered varia- 

 tion in the physiological state of these individuals. That taken 

 by itself, it is interpretable in terms of learning may serv^e as 

 a warning against generalizations based on experiments on one 

 or two individuals. Nevertheless the relation of length of 

 training and training effect in the snail may be worthy of further 

 investigation. 



II. MODIFIABILITY STUDIED BY THE LABYRINTH METHOD 



1. Introduction 



The evidence of learning in the snail, PJiysa gyrina Say, 

 obtained by the use of the method of simultaneous stimuli 

 suggested the query whether the snail could solve a simple 

 labyrinth. Failure to solve the labyrinth would be evidence 

 that this requires a capacity beyond that of forming simple 

 associations. The labyrinth method has been used extensively 



